This article suggests that zero-derivation and apophony were strategies of derivation in pre-Old Turkic. It argues that the semantic features of the Old Turkic noun-verb pairs based on the bare root correspond to the so-called “ergative formation” described in Erdal (1991) for suffixal derivation involving deverbal suffixes, which describes the fact that transitive verbs correspond to object nouns or adjectives and that the subject nouns or adjectives stem from intransitive verbs (Erdal 1991: 169). For instance, the transitive verb ur- ‘to hit, to put’ yields the object noun ur+uɣ ‘seed’, while the intransitive verb kürä- ‘to desert’ plus the same suffix yields the subject noun kürä+g ‘deserter’. This study aims to show that nounverb pairs follow the same pattern, for instance, kes- ‘to cut’ (tr.) and kes ‘piece, part’ (object noun), or ḳarï- ‘to become old’ (itr.) and ḳarï ‘old’ (subject adjective). This implies that non-suffixed nouns are derived from verbs (and not the other way round): toŋ- ‘to be frozen hard’ (itr.) → toŋ ‘frozen’ (subject adj.). Furthermore, this article proposes that derivation through vowel lengthening (apophony) was another form of derivation in Old Turkic, e.g. toz- ‘to become dust’ (itr.) → tōz ‘dust’ (subject noun) and follows the ergative pattern as well. However, cases of zero-derivation or apophony are rather rare, so these derivation patterns might go back to an earlier stage of Turkic.